MCC 172 Minimum Resolution for 100 mV Amplitude Signal Appears to be 10 mV

[APickAluas]

I'm testing an MCC 172 on a Raspberry Pi 400 and have a question regarding the minimum voltage which can be resolved. I acquired a 1 Hz 100 mV signal using a modified version of the finite_scan.py example script where I export the voltage readout to a csv. I'm using a 1 kHz scan rate and recorded 10 k samples. 

My issue is that when plotted the signal appears to only be resolved to 10 mV precision (despite being clearly smooth at a far higher resolution when viewed on a scope instead). I would expect with 24 bits for a -5 to 5 V range the signal would be resolved to ~ .6 μV. Is there something I'm misunderstanding or some option I'm missing in the code to initiate the scan which sets the bit depth lower by default? I'm not sure what else could account for this. Additionally, the amplitude is inaccurate (60 mV instead of 100 mV) when I acquire signals with frequencies lower than ~10 Hz. 

Apologies for the lack of axis labels and units, but here's a plot showing what I've described.

[DAQman]

The MCC 172 inputs are AC coupled, so no DC specification exists. The AC cutoff frequency is 0.78 Hz (-3dB point). This specification and others are listed in the following hardware specification: https://mccdaq.github.io/daqhats/_static/esmcc172.pdf.

Test the dynamic performance with an AC signal above 10 Hz. For example, set your function generator to output a 100 Hz 100mV sine wave. Connect it to a bench meter capable of measuring AC RMS for reference. Then, connect it to the MCC 172 and compare the RMS results from the continuous_scan example. My meter returns 70.794 mV RMS, and my MCC 172 reads the same signal at 70.68 to 70.92 mV RMS.

 

 

 

[APickAluas]

Thanks for the response! I have performed scans with higher frequencies (e.g. 100 Hz which is where we eventually want to be running things anyway) and indeed the amplitude is accurately recovered as 100 mV, but the 10 mV discretization of the signal remains. I'm unclear on what AC coupling could have to do with that?

Here's a plot of the 100 Hz scan:

[DAQman]

You are dealing with a dynamic input for IEPE sensors such as accelerometers. It's not the same as DC input from a typical DAQ device. It uses a 24-bit delta-sigma A/D converter because it can be programmed with a digital filter, which prevents aliasing. The signal-to-noise ratio is one parameter that is typically used to determine performance. There are others, too, such as signal-to-noise and distortion ratio (SINAD) and Total Harmonic Distortion, THD.